Big Four Consulting Faces an AI Dilemma. KPMG's Answer Is Claude.

Walk into any strategy meeting at a Big Four firm and two anxieties fill the room. The first is that AI tools will produce wrong answers with dangerous confidence, and that the first firm to bet heavily on one will inherit the liability when a deliverable turns out to be wrong. The second fear runs almost exactly opposite: that the AI companies keep improving, that they start packaging their own industry expertise, and that one day a CFO decides to skip the consulting firm entirely and go straight to the model. KPMG's sweeping Claude commitment, announced last week and analyzed in detail by Fortune on Monday, is the firm's attempt to resolve both problems with a single move.

The mechanics of the deal are by now well documented. Every employee across all 138 countries where KPMG operates gets access to Claude. The firm is embedding Claude directly into Digital Gateway, its proprietary client delivery platform, making it the first of the four major accounting and consulting firms to put AI at the center of its client-facing infrastructure rather than treating it as an add-on. KPMG is also becoming Anthropic's preferred consulting partner for private equity, meaning it will help PE firms deploy AI inside their portfolio companies, a segment that sits at the intersection of KPMG's advisory work and Anthropic's rapidly expanding enterprise ambitions.

The Hallucination Problem

The reliability fear is real and not confined to KPMG's boardroom. In professional services, a factual error in a deliverable can expose a firm to malpractice claims, regulatory scrutiny, and the kind of reputational damage that takes years to repair. Early generative AI tools produced authoritative-sounding text that turned out to be fabricated. That history has made senior partners at every major firm cautious about how quickly to put AI in front of clients.

KPMG's bet is that Claude's accuracy on professional workflows has crossed a threshold where the productivity gains outweigh the reliability risks, and that building internal familiarity at scale is the only way to know where the edges are before a client finds them. The exclusive KPMG arrangement for private equity work is also a signal: Anthropic is willing to constrain which firms can bring Claude to that market, which implies a confidence in the outputs that a commodity API relationship would not support.

The Disintermediation Fear

The second anxiety is structurally different, and in some ways harder to address. Consulting firms derive value from accumulated domain expertise: the partner who has seen forty M&A integrations knows which ones failed and why. That knowledge does not sit in a database; it lives in the heads of expensive senior professionals. If an AI model can synthesize that expertise from training data and offer it at a fraction of the cost, the business case for the consulting firm weakens.

Anthropic has done nothing to hide its enterprise ambitions. The company's partnership with Blackstone and Goldman Sachs for the midmarket AI services venture announced earlier this month is explicitly designed to take AI directly to businesses that currently rely on consulting firms to act as integrators. PwC is training 30,000 professionals on Claude, building a Claude Code and Cowork practice that will compete for the same enterprise budget. The Big Four are, in effect, financing the expansion of a technology that could eventually undercut their own fee structures.

KPMG's response is to make the bet legible. Rather than hedging across multiple AI vendors and hoping none of them become direct competitors, the firm has chosen Anthropic as a close partner and taken an exclusive position in a high-value market segment. The logic is that a firm deeply embedded in how Claude works, and with a preferred-partner agreement that gives it early access to new capabilities, is better positioned to retain clients than one that treats AI as a commodity input.

"Every employee, in every one of the 138 countries where KPMG operates, will now have access to Claude." Fortune, May 26, 2026

What This Means for the Rest of the Big Four

KPMG's move creates pressure on its competitors. Deloitte, EY, and PwC all have AI programs, but none has yet made a commitment at this scale or depth. PwC comes closest, with its expanded Anthropic alliance and the 30,000-professional certification program. But PwC has not embedded Claude into its delivery infrastructure the way KPMG has with Digital Gateway, and it does not hold the private equity exclusivity.

For clients, the competitive dynamic may ultimately be favorable. If multiple Big Four firms are racing to put the best AI capabilities in front of their engagement teams, the output quality of consulting work should improve faster than it would if firms adopted AI cautiously and gradually. The risk is that the race creates uneven reliability: a firm that moves fast learns where the failure modes are, but so do its clients.

The broader question, which neither KPMG nor Anthropic has fully answered, is where the line sits between AI-assisted consulting and AI-delivered consulting. Right now, the model is that Claude helps KPMG professionals work faster and more consistently. Whether, over time, clients start to see Claude as the primary value and the KPMG brand as the delivery wrapper is a question that will take years to resolve. For now, the direction of travel is clear: the Big Four have decided that the right response to AI disruption is to be closer to the AI, not farther from it.